The purpose of this study was to examine the potential effect of false assumptions regarding the motivation of examinees on item calibration and test construction. A simulation study was conducted using data generated by means of several models of examinee item response behaviors (the three-parameter logistic model alone and in combination with Wise's Examinee Persistence model). Items were calibrated using a marginal maximum likelihood estimation method with an expectation/maximization algorithm with Bayesian prior distributions for item parameter estimates. Results clearly document the possible effect on item calibration of examinees who display low motivation response behaviors. For the conditions studied, item parameter estimates based on responses of examinees in low motivation conditions were more biased and more variable than items calibrated using responses from a calibration group who behaved according to the three-parameter logistic model.